Multi-Agent, Interaction, Décision

Computing models designed by the MIND Team are integrated into software agents interacting with users, into multi-agent systems in which agents and humans coexist, in particular in mixed communities, or into cyber-physical systems.

The main aspect of the MIND team's research consists in placing users at the heart of its approach, from data collection and processing (using machine learning and graph analysis algorithms), knowledge elicitation, experimental model validation, to experience optimization as decision-making tool. The aim is to develop integrated thinking processes in which automated decision steps interact with human input or with human users, potentially guided by software agents.

The challenge of each framework is to achieve the appropriate balance between user control required by acceptability, responsibility or by the decision-making context.

The objective also includes the delegation to software agents in order to reduce the user cognitive load. 

Case recognition and agent decision-making are based on reasoning patterns arising from distributed artificial intelligence and semantic technologies covering formal knowledge modeling and several types of reasoning (temporal, spatial, fuzzy, probabilistic, etc.) Experience capitalization and meta-knowledge approaches are also applied to optimize case recognition and decision-making processes. 

Lastly, specific interactions between user and software agent must be considered from both user and agent angles.

The aim is to pick up, represent and analyze messages and signals from users appropriately, in order to tailor a software solution (e.g. personalization of system behavior) or to trigger warnings (e.g. privacy violation risk).

This means matching the communication method with the user and the decision-making context, both in terms of content (identification of relevant and reliable information) and in terms of design (ecological interactions through embodied chatbots and emotional communication, etc.).

    • Embodied chatbots and mixed human-agent communities

    • IT environment/Digital learning

    • Social networks

    • Customized information retrieval

    • Industry 4.0 and cyber-physical systems

    • Internet of Things